Modeling Open Data Usage: Decision Tree Approach

Abstract

Predicting the quality of datasets is important for individuals, private or public organizations regardless of their intent, especially in today’s expeditious and competitive environment. If data are considered as valuable, then data need to have a stronghold in their quality or rather in indicators from which the quality consists of. Although there are several data mining methods that can be used for this issue, the regression tree method is used because of its advantages in comparison with other methods. Some of the pros are that tree is very easy to understand or explain, it is a collection of if-then rules, it can be used for investigating relationship between predictors and target attribute without understanding the form of the relationships. In this paper, open data usage predicting task was performed on the data collected from open data portal Data.gov.uk. Results show that it is possible to build a model of high accuracy.